yxu0611/tensorflow-implementation-of-lcnn

A Tensorflow implementation of "A Light CNN for Deep Face Representation with Noisy Labels"

tensorflow-implementation-of-lcnn's Issues

how to train MS-celeb-1M?

Appreciate for your code.

I want to train LCNN, there are no information for training In the readme.md.
Could you help me?

Training LightCNN-29 with MsCeleb-Faces-Aligned

When Training LightCNN-29 with MsCeleb-Faces-Aligned, we can see: the training loss decreases slowly, and the training/valing accuracy increase also slowly.
Could you give some advises to speed the training phase?
For example, the initialization methods for init the weights, the learning rate scheduler, and so on.
Thanks

Image preproccessing

Hi
How can I align and crop the training dataset ?

file.txt

How to save the dataset for CASIA WebFace in a list.txt such as lfw_patch_part.txt for the case of the lfw dataset

pretrained model download

I can not download the model from onedrive ?can anyone share me one in baidunetdisk or googledrive? Thank you!

Could you please provide the imglist_80k.txt ?

Could you please provide the imglist_80k.txt ?

Is MFM_fc1 missing?

Hello,

Thank you for the effort to port it to Tensorflow.

In the original paper fc1 is followed by MFM_fc1. I see you implemented MFMfc but it's not used anywhere. Is it a bug or intentionally made decision?

about accuracy

I wan to train a Feature Extractor use your model，When should I stop training?
I use your model the accuracy in training set is about 80% no longer improve ，so what should I do?

Welcome to raise issues and your requests here

Is there any pretrained model provied?

:P

about accuracy

I wan't to train a Feature Extractor use your model，When should I stop training?
I use your model the accuracy in training set is about 80% no longer improve ，so what should I do?